Moore's Law Staying Strong Through 30nm

on the BUSS by opencity · 2006-02-20 02:31 · Score: 4, Interesting

At what point does BUSS technology break down? Figured this was where to ask.

--
Physics is like sex: sure, it may give some practical results, but that's not why we do it.

Re:on the BUSS by RabidMoose · 2006-02-20 03:30 · Score: 4, Informative

To whoever modded the parent redundant. It was the first post. How is that possible?
Re:on the BUSS by Anonymous Coward · 2006-02-20 03:33 · Score: 0

spelled with one s
Re:on the BUSS by default+luser · 2006-02-20 08:57 · Score: 1

On copper, the speed is probably limited to somewhere in the 10 GHz range even for short drops, much lower if you have a parallel bus. But given serial interfaces like HyperTransport, you can just keep adding lanes (to a reasonable limit) to increase bandwidth.

When we reach the limit of copper serial busses, we branch out to optical serial busses, which have the potential to run as fast as hundreds of GHz. Probably won't see these for at least a decade.

--
Man is the animal that laughs.
And occasionally whores for Karma.

What's the minium then? by bronney · 2006-02-20 02:31 · Score: 4, Interesting

I am too lazy to learn these things from scratch but would anyone cared to tell us what's the theoretical minimum width we can go before eletrons starts jumping wires? I hope it's not 5nm.

Re:What's the minium then? by lbrandy · 2006-02-20 02:55 · Score: 1

I am too lazy to learn these things from scratch but would anyone cared to tell us what's the theoretical minimum width we can go before eletrons starts jumping wires? I hope it's not 5nm.

The theoretical minimum width for the current type of transistors is in the .2nm range, the width of a single atom. Tunneling and other quantum effects will very likely prevent us from ever getting that low, however.
Re:What's the minium then? by corngrower · 2006-02-20 03:29 · Score: 1

Electrons may not be jumping wires, but they're already starting to jump across
the gate dielectric. This is only a few (dozen or two) atomic layers in thickness already.

Lithography is not the only problem that must be solved in order to improve
the density of the chips. There are problems involving the gate oxide, the
dielectric of the insulator, routing, leakage currents, and interconnect
capacitance issues.

The chips may get more dense, but the individual gates may no longer be getting
faster. Getting faster chips will increasingly be dependant on implementing parallelism or pipelining at higher levels. Chip design will move away from
'Von Neuman' (or derivitive) type processor architectures.
Re:What's the minium then? by schlpbch · 2006-02-20 04:27 · Score: 0, Offtopic

Come on, it's not that diffcult to learn, especially when it's explained by such a charming teacher:

http://britneyspears.ac/lasers.htm/

I've heard that one before... by Aslan72 · 2006-02-20 02:34 · Score: 3, Insightful

"This buys Moore's Law a few more years."

I've heard that more than a few times.Isn't that why it's a law? It seems like every 18 months or so, Moore ends up almost petering out (kind of like apple...) and there ends up being a redeeming breakthrough that keeps it around.

If it wasn't a law, we'd just call it Moore's hypothesis, or Moore's pittiful attempt at justifying an upgrade. I remember the day when 50Mhz was the theoretical limit for speed and then they got the grand idea of putting a heat sink on the chip.

--pete

Re:I've heard that one before... by Waffle+Iron · 2006-02-20 02:48 · Score: 5, Informative

Isn't that why it's a law?
It's not a law. It's just incorrectly called a law.
It should be plainly obvious that any exponentially increasing phenomenon can't be a "law". If this so-called law were to continue unabated for a couple of centuries, the number of transistors in a chip would exceed the number of atoms on planet earth. Clearly, a limit is going to be reached well before that happens.
Re:I've heard that one before... by OverlordQ · 2006-02-20 02:51 · Score: 4, Funny

yes, by then we will all have Nural Net Processahs.

--
Your hair look like poop, Bob! - Wanker.
Re:I've heard that one before... by Nikker · 2006-02-20 03:09 · Score: 0

You know the funny thing is that Moore was the co-founder of Intel and sais that evrey 18 months chip speed will double. The question is what is the factor that makes it take 18 months? I find myself doubtful that Intel brain power sits around for 16 months comming to the conclusion that making the process smaller is the way to go.

If the company came to the correlation that size effects speed do you think they would really start at a process 10um and restrict themselves to small advances at a time? I know if I was in their position I would definately figure if speed = small size I would start looking at the smallest off the bat. If so then 18 months is likely the time it takes to make enough off of the last batch to afford the new ones.

--
A loop, by its nature, continues. If that didn't make sense, start reading this sentence again.
Re:I've heard that one before... by Directrix1 · 2006-02-20 03:12 · Score: 4, Funny

You obviously have not had enough spelling training data supplied to your neural net.

--
Occam's razor is the blind faith in the natural selection of least resistance and in universal oversimplification. -- EF
Re:I've heard that one before... by Aslan72 · 2006-02-20 03:16 · Score: 1

And at 3:02 a.m. CST on April 8th, 2037 Windows became self-aware.
Re:I've heard that one before... by Cobralisk · 2006-02-20 03:28 · Score: 2, Funny

Theah is sahmthing wrong with youah sense of humah circuit. Perhaps You need more powah!!!

--
Waiting for ad.doubleclick.net...
Re:I've heard that one before... by Bad+to+the+Ben · 2006-02-20 03:33 · Score: 2, Insightful

That's kind of like saying the Wright Brothers should have skipped the Wright Flyer and just built a 747. It doesn't work that way, it's not that simple.

Gearing up for a processor run can't just be done overnight. A fabrication line has to be created, and chip designs crafted to build certain chips at a certain process. This takes time (ask AMD). Whilst this is being done, the next process would be being researched, and ways would be discovered to make the new process profitable and not ridiculously expensive. Then, to build chips at the new process, new chip designs and fab lines have to be done. This
takes time. Whilst this is being done, the next process would be being researched, etc etc. If Intel perpetually waited for "the smallest possible process" we'd never get any chips.

There probably is an element of truth to your argument, I'm sure Intel does try to milk the most out of it's existing run to benefit from economies of scale. But scaling to the next process is not a simple task.
Re:I've heard that one before... by QuantumFTL · 2006-02-20 03:33 · Score: 1

It should be plainly obvious that any exponentially increasing phenomenon can't be a "law". If this so-called law were to continue unabated for a couple of centuries, the number of transistors in a chip would exceed the number of atoms on planet earth. Clearly, a limit is going to be reached well before that happens.

Unless you use something smaller than atoms...
Re:I've heard that one before... by metarox · 2006-02-20 03:34 · Score: 1

What you miss here is that the companies will milk the cow at every level they can to make their life longer. I'm pretty sure they could just jump down to 45nm or even 30nm right away, but hey, if we do that, after everyone has the best, nothing more to buy.
Re:I've heard that one before... by xornor · 2006-02-20 04:03 · Score: 2, Funny

and 3 nanoseconds later killed itself.
Re:I've heard that one before... by RhettLivingston · 2006-02-20 04:19 · Score: 3, Interesting

Sure, its not a law. But...

I'm not sure that it is so clear that the limit will truly be reached before a processor capable of performing as if it had a transistor for every atom of the earth is created. Assuming we're still around, I believe we'll be able to maintain the increases in speed and scale predicted by Moore's law through means we can only just imagine now.

Certainly, it is starting to appear that we'll see combinations of quantum and other processing. There was also recently a development in tri-state per bit quantum storage that may be extendable to n-state per bit. Perhaps we'll find ways to put subatomic particles together into things other than atoms that don't even require atoms as a trapping mechanism and be able to fully exploit that scale. We could explore processing in ways where a single "transistor" or whatever happens to be the smallest scale component participates in different ways in multiple operations or memories like neurons already do. Technologies for processing that don't generate anywhere near as much waste heat are appearing (magnetic for instance) thus allowing the full exploitation of the third dimension to look more plausible without hitting heat dissapation barriers (solid cubes instead of layered wafers). And what about other dimensions? At the atomic scales we're reaching, it is much more believable that we'll eventually be able to exploit some physical phenomenon to put some of the processing or storage mechanisms into non-temporospatial dimensions.

Anyway, I believe it to be very unimaginative to say that Moore's Law will ever hit a barrier. I would call it a virtual law. Sure, its not a "law" as in a law of physics. It isn't a theory either. Rather, its a good guess at a rate of development that we can sustain.

I personally believe that the law is going to change in a few more years as computers reach a level of sophistication necessary to directly participate in more of the scientific research necessary to bootstrap the next generation, gradually eliminating the man in the loop unless we find ways to start scaling the brain's capabilities. At that point, we may start to see the 18 months per generation become one of the variables of the law that is scaling down toward 0.
Re:I've heard that one before... by Waffle+Iron · 2006-02-20 04:24 · Score: 1

Unless you use something smaller than atoms...
Generally, the smaller the particle, the higher the mass and/or energy of that particle. Neither one of those properties seem to be helpful for trying to create high speed or low power circuits.
So even if you could find a stable form of matter with a much denser structure than atomic matter, and even if this matter could be safely handled (although I suspect that the most likely application of such matter would be for WMDs or interstellar rocket fuel), it's not clear that you could use it to create high-performance logic circuits.
Re:I've heard that one before... by Surt · 2006-02-20 04:32 · Score: 1

It is a law:

1 a (1) : a binding custom or practice of a community : a rule of conduct or action prescribed or formally recognized as binding or enforced by a controlling authority

That's the very first definition of law from m-w.

So in particular, the controlling authority (Moore) has decreed that transistor density shall double at such and such a rate, and the industry has obeyed this rule of conduct.

--
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
Re:I've heard that one before... by databyss · 2006-02-20 04:37 · Score: 0

First off, doubling is not an exponential increase.

Secondly, I think that before they get to any insane material limits, they'll be referring to the Equivalent Transistor Speed or future processors, by which Moore's Law might still hold.

--
Hmmm witty sig or funny sig? Maybe elitest techy sig!
Re:I've heard that one before... by trentblase · 2006-02-20 04:40 · Score: 1

Generally, the smaller the particle, the higher the mass and/or energy of that particle.
Dumbest. Statement. Ever. Unless you meant density. Then I'd believe you.
Re:I've heard that one before... by trentblase · 2006-02-20 04:42 · Score: 2, Insightful

First off, doubling at a constant rate is a perfect example of exponential behavior.
Secondly, Moore's law is about transistors per chip, so maybe you mean Equivalent Transistor Count.
Re:I've heard that one before... by hackstraw · 2006-02-20 04:53 · Score: 1

It's not a law. It's just incorrectly called a law.

Right!

It should be plainly obvious that any exponentially increasing phenomenon can't be a "law".

Wrong!

For those that don't know, Moore's "law" says something like, "Every 18 months, humans (hopefully those that work for Intel) will figure out how to double the existing transistor count in a CPU".

Anybody that can put an open ended and exponentially increasing assumption on human behavior _AND_ call it a law should be in marketing, not science.

Now, regarding CPU performance and whatnot, am I the only one who believes that performance has not significantly increased over the past 3 years (2 Moore cycles)? Sure, they may be able to put more crap in the CPUs, but they surely haven't seemed to get any faster.
Re:I've heard that one before... by RicktheBrick · 2006-02-20 04:57 · Score: 2, Interesting

My question is this "What are we getting out of these faster and more powerful computers?". There have been several new super computers brought online in the last few year. Many of these are capable of doing over 50 trillion calculations a second. Yet when President Bush asks this country to develop technology to reduce our need for foreign oil, he says we need 20 years to do so. On a personal note I want a small bed that I can totally shut off the rest of the world. I want it to be totally sound proof and be the same temperature and humidity all the time. In order to do this the bed must be small so that I do not have to spend too much money to maintain it and I must have a computer that will take care of the rest of the house so that I will still be alive the next morning and will find the rest of my house still intact. In order to do this the computer will have to be able to listen to all the noises in the house and determine if that noise will need my assistance to either correct it or flee from it. I want a computer that I can trust my life with.
Re:I've heard that one before... by Anonym0us+Cow+Herd · 2006-02-20 05:02 · Score: 1

Unless you use something smaller than atoms...
Smaller than atoms is okay. Just don't get smaller than the Plank Length.

--
The price of freedom is eternal litigation.
Re:I've heard that one before... by Waffle+Iron · 2006-02-20 05:23 · Score: 1

Please go learn something about quantum mechanics before you post your judgements.
Re:I've heard that one before... by SatanicPuppy · 2006-02-20 05:32 · Score: 1

They always say stuff like this. The old systems were limited because you could never have more than x number of vacuum tubes. Then it went to chips, and there was always some reason why the chip couldn't get any bigger or denser or hotter...and there always turned out to be another way.

So now we can't have more gates than there are atoms...but what if we're using subatomic particles, so that one atom's worth of particles can comprise multiple gates? What if we find some way to move beyond gates, so that we have an increase in capability coupled with a decrease in mass?

I'm not arguing for the perpetual truth of Moore's Law, but don't fall into the trap of assuming that future techniques are going to be limited by our current understanding.

--
ad logicam Claiming a proposition is false because it was presented as the conclusion of a fallacious argument.
Re:I've heard that one before... by RhettLivingston · 2006-02-20 05:35 · Score: 1

First, be careful what you wish for. You'd better add some sensory data to that sleep environment you claim to want. Sensory deprivation is an extremely effective means of driving someone insane. You may be a vegetable before you get through the first night :-)

Second, be careful what you wish for. It seems that you're wishing for a computer with extensive capabilities of gathering data about you and your environment, the capability of making complex judgements concerning whether your safety or a probable multitude of other things that concern you has crossed some threshold requiring action, and the capability to act on that when it is in your best interests. Hmmmm.

Depending on which extreme of the estimates you believe, we will have computers with the processing and storage potential of the human brain in something like 9 to 19 years. Of course, we're unlikely to know how to program it. We'll solve that problem by figuring out how to download an existing program to it, i.e. scan an existing brain's program. That may take 5 years if they aren't already working on it. Give 5 years past that for the supercomputer initially needed to scale down to a chip (will happen at an accelerated rate because the first few people scanned into these systems will be experts in microelectronics, physics, etc. and their capabilities will scale somewhat as the machines scale), and you'll have the full potential for your gadget. So, a plausible answer for you is sometime between 2020 (assuming some development overlap) and 2035. But, I don't wish for this development.
Re:I've heard that one before... by zardor · 2006-02-20 05:37 · Score: 2, Funny

Go easy on him. He's just running on a single core....

--
-- We don't understand software, and sometimes we don't understand hardware, but we can *see* the blinking lights
Re:I've heard that one before... by anorlunda · 2006-02-20 06:53 · Score: 1

What a remarkable law it is. For thirty years I've been reading that development will soon hit a brick wall and Moore's Law will end. A corollary to the nay saying is that silicon technology will have to give way to something new; everything from silicon on sapphire, to organic molecules. Hasn't happened yet. Moore's Law and silicon still rule in this world.

Even more remarkable than Moore's prediction that growth would be exponential, is the fact that he nailed the actual value of the exponent. He only had to make one adjustment in the exponent's value since the beginning.

I read an article that says that it's not entirely an accident. The chip fabricators also believe in Moore's law and they make their business plans and their R&D plans based on it. They fear that they can't stay competitive if they fall behind Moore's law.
When they succeed in their goals, they cause Moore's Law to be fulfilled almost exactly. Thus, Moore's Law is partially self-fulfilling and Mr. Moore not only foretold the future, he shaped it.

My hat goes off to Gordon E. Moore. He should be awarded the Nobel Prize in Physics for his contribution.
Re:I've heard that one before... by corngrower · 2006-02-20 06:59 · Score: 2, Informative

Well Moore's law is almost petering out again. Some serious difficulties appear to be cropping their heads once you go below about 30nm, not that there aren't substantial difficulties already.
Interconnect capacitance is starting to be a real killer. As transistor sizes shrink, their capacity to source & sink current drops a bit. Even with using copper for the interconnect layers, because the cross section of these wires is so small the resistance is non negligible. What this all means is that the time required for the signal to travel over the wires that interconnect the transistors is on the average increasing even as feature size decreases. With gigahertz clock speeds, it already takes several clock cycles for a signal to travel from one side of a large chip to the other. Buffers have to be inserted along wires every several micrometers to keep signal delays from increasing quadratically with length.
The material that is used for gate dielectrics, currently silicon dioxide, will likely have to be different. Something with a higher dielectric is needed. Problems are starting to occur because this oxide layer is getting so thin.
Re:I've heard that one before... by FatAlb3rt · 2006-02-20 07:16 · Score: 1

So it buys us a few more years...I'm just wondering what's been happening the last 2 or 3 - it seems we've been stuck with the 3 GHz processors for a little too long now.
Re:I've heard that one before... by Anonymous Coward · 2006-02-20 07:32 · Score: 0

>It seems like every 18 months or so, Moore ends up almost petering out (kind of like apple...) and there ends up being a redeeming breakthrough that keeps it around.

We had 3.4 GHz processors in 2001, and we have 3.4GHz dual core processors now. As far as I can tell the "redeeming breakthrough" was adding multiple cores? If so, they should be selling 8-core 3.4 GHz processors by now.

Moore's Law is dead. R.I.P.

p.s. What's really sad is: since they're not doubling processing capability every 18 months, we're no longer on schedule for a computer that can simulate the human brain (neuron for neuron) by 2027. That means "true AI" is probably at least 50-60 years off instead of 25-30. (* if it's even possible to have a computer that powerful.)

p.p.s. What's really cool is that my bargain bin 2.4 GHz processor from 2001 is still reasonably fast by modern standards, so there's no need for me to consider upgrading. Any oldtimers out the remember buying a brand new Pentium 90-100 and thinking it was awesome? :)
Re:I've heard that one before... by Doppler00 · 2006-02-20 07:47 · Score: 1

You have to realize, these are the same people in Silicon Valley who have seen their real-estate values go up 20% year over year for the last five years. They like believing that things only go up exponentially without any leveling off.
Re:I've heard that one before... by trentblase · 2006-02-20 07:55 · Score: 1

Enlighten me. None of the particles I've looked up seem to work that way.
Re:I've heard that one before... by Wyatt+Earp · 2006-02-20 08:14 · Score: 1

Be hard for us to have a computer that has the processing and storage potential of the human brain when those aren't known.

While Hans Moravec guestimates, by extrapolating from known capabilities of the retina to process image inputs, a brain has a processing capacity of 100 trillion instructions per second, and is likely to be surpassed by computers by 2030. http://www.transhumanist.com/volume1/moravec.htm

That is a guestimate at best.

Furthermore the Brain unlike a computer is able to do amazing things even when it has suffered terrible damage, see Kim Peek
http://en.wikipedia.org/wiki/Kim_Peek

I'd go on, but I'm off to the Doctor about my Stroke damaged brain :)
Re:I've heard that one before... by Waffle+Iron · 2006-02-20 08:16 · Score: 1

The Heisenberg uncertainty principle states that the lower the mass of a particle, the greater the uncertainty in its position. The "size" of atoms is determined mainly by the uncertainty in position of the low-mass electrons that surround them. (The nucleus of an atom is far smaller than the atom as a whole because the protons and neutrons in the nucleus are orders of magnitude more massive than the electrons.)
To create matter with stable features packed more densely than ordinary atoms, you'd have to use only particles that weigh more than electrons. Otherwise, the uncertainty in their positions would inflate their effective sizes to be just as big as what we work with today.
Examples of such matter have been created. For example, replacing an electron with a negatively charged muon (which is ~100 times heavier than an electron) in a hydrogen atom creates a very compact atom. However, muons are unstable and these atoms live only a few microseconds. (That may be a good thing since the compactness also allows the atoms to undergo nuclear fusion at room temperature, so a bunch of that stuff in one place would be an extremely dangerous and volatile mixture.)
Re:I've heard that one before... by databyss · 2006-02-20 08:29 · Score: 1

Quite true. I am very dumb today, for that I apologize.

And the Equivalent Transistor Count is what I had intended to write yet somehow my brain garbled my thoughts.

That makes the score: Me=0 Intarwebs=2 for today.

--
Hmmm witty sig or funny sig? Maybe elitest techy sig!
Re:I've heard that one before... by AaronPSU777 · 2006-02-20 08:31 · Score: 1

Silver has superior electrical conductivity to copper. The delta is not as big as it was from aluminum to copper but it is certainly there. I wonder what the roadblocks are preventing its use in interconnects.
Re:I've heard that one before... by kula.shinoda · 2006-02-20 08:35 · Score: 1

Who said that future chips have to have transistors in them? You don't know what technology will bring.

--
Real men don't write sigs
Re:I've heard that one before... by Anonymous Coward · 2006-02-20 08:41 · Score: 0

assuming it was loaded on a contemporary machine and following moores law, it would be running at a smidge under 5 PHz... imagine the damage it could do in 3ns
Re:I've heard that one before... by Just+Some+Guy · 2006-02-20 08:41 · Score: 1

If this so-called law were to continue unabated for a couple of centuries, the number of transistors in a chip would exceed the number of atoms on planet earth.
Quite a few very intelligent people think that's exactly what will happen, possibly within our (extended) lifetimes.

--
Dewey, what part of this looks like authorities should be involved?
Re:I've heard that one before... by Waffle+Iron · 2006-02-20 09:15 · Score: 1

I don't care what the chips have in them. If you don't have enough fundamental particles to build the features, Moore's law must come to an end. By my math, with 18-month doubling, the chip feature count exceeds the number of particles in the known universe in less than 400 years.
Re:I've heard that one before... by RhettLivingston · 2006-02-20 09:48 · Score: 1

That's exactly the reason for the range. The last time I researched this (been about a year) the general years when we'd cross the threshold were between 2015 and 2025 depending on how much capability the various paper writers estimated that the brain had while assuming Moore's Law as a constant. 2030 is a bit outside the norm of what I was seeing at that time, but what the hay, add five to my upper estimate and push it out some.

Your other issues are primarily a programming issue, which is why I indicated that we'll have to be able to scan the "program" out of an existing brain. I don't see us ever being able to truly "understand" a program this complex but developers will find a way where researchers fail on this one. I did leave out the remapping of that program onto a massively parallel hardware architecture. That will take some time because all of the delay pipelines must be maintained as a critical part of the program.

In any case, we're more likely to hit this point in the first half of the century than the second half. So, many of us stand a good chance of seeing it barring catastrophes.

Your example actually brings up another interesting research area. Those who have been reverse engineering the optic path to an accurate analog model have found that a tremendous amount of vital processing is performed in the optic nerve path prior to the signal reaching what we generally consider to be "the brain". I've seen overestimates of brain function occur from this. The brain is processing a sensor signal that already contains edge, movement (extracted by splitting signals into multiple paths with different time delays and then passing them through comparators), and lots of other information. So much abstraction has already occurred that the pixelated type of image that most of us think of processing doesn't exist at the brain level. I think that hardware effort was occurring at Stanford if anyone cares to track it down. Cool stuff.
Re:I've heard that one before... by JonathanBoyd · 2006-02-20 10:37 · Score: 1

The Heisenberg uncertainty principle states that the lower the mass of a particle, the greater the uncertainty in its position.

No it doesn't. It states that the product of the uncertainty in the momentum and the uncertainty in the position is equal to or greater than Planck's constant divided by 4 pi. That can mean in certain situations that what you said is true, but it is dependent on the velocities of the particles.
Re:I've heard that one before... by kula.shinoda · 2006-02-20 11:03 · Score: 1

Well of course you're correct, from our current point of view. I see no reason why future science may yet bring better results. 400 years of science is a long time, in which we could find ways around these things.

My argument may seem philosophical, but get back to me in 20 years :)

--
Real men don't write sigs
Re:I've heard that one before... by Waffle+Iron · 2006-02-20 12:08 · Score: 1

In the context of a stationary solid material as I was discussing, it's true enough.
Re:I've heard that one before... by poopdeville · 2006-02-20 15:20 · Score: 2, Informative

Doubling is an exponential increase. http://en.wikipedia.org/wiki/Exponential_function

--
After all, I am strangely colored.
Re:I've heard that one before... by Zeinfeld · 2006-02-20 16:30 · Score: 1

Generally, the smaller the particle, the higher the mass and/or energy of that particle. Neither one of those properties seem to be helpful for trying to create high speed or low power circuits.
Protons weigh less than a hydrogen atom but they are much smaller. So no does not seem to fit.
Exotic quarks like top, botom, beauty, charm weigh much more than ordinary up or down quarks. But they do not have a meaningful size.
A much bigger obstacle to using any form of exotic particle in a chip would be the fact that 1) they are exhorbitantly expensive to make, all the exotic particles created since particle physics started would fit comfortably in a ball the size of a pea, 2) they are unstable lasting tiny fractions of a second.
We are probably closer to making a warp drive work and certainly a lot closer to making a fussion reactor than making ICs out of short lived subatomic particles.

--
Looking for an Information Security student project suggestion?
Try http://dotcrimeManifesto.com/
Re:I've heard that one before... by trentblase · 2006-02-21 06:16 · Score: 1

So you're saying that an electron is "bigger" than a neutron because it has a greater uncertainty of position? That's a poor defintion of size.
Re:I've heard that one before... by Waffle+Iron · 2006-02-21 06:47 · Score: 1

Nevertheless, that's how it works once you start piling them together in bulk. The sizes of the electrons' orbitals, which are determined by their probabilistic wave functions, largely define the sizes of the atoms.
Re:I've heard that one before... by xpyr · 2006-02-21 09:04 · Score: 1

Moore ends up almost petering out (kind of like apple...) and there ends up being a redeeming breakthrough that keeps it around.

Yeah, peopole always say apple is on the brink of going out of business. But it survives because it has higher profit margins and its occult following of loyal mac zealots. Apple fits well in an ecosystem where it has the main monopoly. But the moment actual competition comes in, it starts to lose out. It's like it doesn't know how to compete and lives off its own arrogance. In order for it to survive, it has to have a unique product that no one else can offer and that competitors can't use the same exact technology at a cheaper price. That's part of the reason why apple pantented its scroll wheel on the ipod. That keeps competitors from putting that feature on their mp3 players. Same thing with the drm that the itunes music store has.

Apple can only thrive in an environment where no other product can match its products for features.

--
My Gawd WTF...

Glad to see IBM catching up... by merced317 · 2006-02-20 02:36 · Score: 5, Informative

since RIT has been doing 26nm. http://www.physorg.com/news10755.html

Re:Glad to see IBM catching up... by Alioth · 2006-02-20 02:41 · Score: 2

TFA doesn't say, but perhaps IBM are developing the industrial process (which will usually come after the initial research that says it's possible, which is probably what RIT were doing). It's one thing to show something is possible in a lab, and another thing to develop a process to do it on a large scale.

--
Oolite: Elite-like game. For Mac, Linux and Windows
Re:Glad to see IBM catching up... by dwbassett42 · 2006-02-20 03:09 · Score: 1

Nah, nobody has done anything more at the ~30nm scale than simply creating interference patterns to create lines and spaces. But I'm attending that conference this week, so I'll be sure to check out thier keynote address.

RIT does have a good immersion tool, we (UT at Austin) went over to RIT to use thier tool to test some immersion fluids we were looking at. We put in our high-index fluid, and we had 32nm spaces and lines just like that!

But beyond that, getting an immersion fluid with an index of refraction significantly higher than water is full of problems. Adding ions to water won't work because it makes the fluid too highly absorbing at 193. Organic fluids are the only other available choice, but they have very low surface tension and high viscosity, which will make scanning speeds much lower and therefore the prices of devices made with them much more expensive.
Re:Glad to see IBM catching up... by stevesliva · 2006-02-20 04:38 · Score: 2, Insightful

Your link doesn't actually explain what EWL is, but it's probably a reasonable assumption to assume that it won't be very compatible with 193nm (light wavelength) litho equipment.
IBM's annoucement has a lot to do with stretching the usefulness of existing litho equipment and materials down to nodes that it was never expected to reach. This has been done again (65nm) and again (45nm) from what was once expected. IBM is saying, add water, and we'll do it again (30nm).

--
Who do you get to be an expert to tell you something's not obvious? The least insightful person you can find? -J Roberts
Re:Glad to see IBM catching up... by Anonymous Coward · 2006-02-20 06:57 · Score: 0

But not at RIT. They loooove to brag.

Why Good news for Intel? by Anonymous Coward · 2006-02-20 02:37 · Score: 1, Interesting

Why not good news for IBM? Lets look at what IBM and Intel are really doing now with processors. Power.org is a group formed by IBM is that the Power ISA and such are more standardized between companies and the companies involved with Power.org actually has some say in the next generation processors from IBM (and Freescale, yes Freescale is again in relationship with IBM unlike Apple). Now look at Intel, nothing, no input, nobody can just go up to Intel and ask for feature XYZ without first going through a lot of lawyers. Even AMD's ISA is not the same as Intel's and it is getting worse.

IBM invented this, so IBM should be able to use it more than Intel ever can. I will note that Apple left IBM/Freescale to goto Intel but there are more than just processors reasons why they left. Apple did not from the looks of it feel that IBM was going in the right direction though IBM is going to be better off without Apple involved because Apple has ego issues.

Overclocking? Really? by Anonymous Coward · 2006-02-20 02:38 · Score: 1, Interesting

Is this really good news for overclockers? Some overclocked P4s have already failed due to electron migration. Isn't a shift down to 30nm just going to exacerbate the problem?

Punctuated Equilibrium - Phase Transition in Moore by QuantumFTL · 2006-02-20 02:39 · Score: 4, Interesting

I believe Moore's Law (or, rather, the modified version about processor speed rather than transitor count) will transition to a new regime soon - that of "average" exponential improvement in the form of a punctuated near-equilibrium.

I believe that the chip industry will have to shift paradigms as the limit of a technology approaches and during these shifts there will be a period of relative nonimprovement as new techniques are refined, implemented, and large scale facilities are built.

There's so many promising technologies on the horizon (photonic computing, three dimensional "chips," quantum computation) etc, but the transition to each will be very bumpy, not at all smooth like the last 40 years of refining two-dimensional semiconductors.

As times change, what we know as Moore's law will change with it. It's likely that the "average" improvement will continue to follow the law more or less (considering that it is driven more heavily by economics than technology). Computers will continue to get faster, cheaper, and able to do things we wouldn't have thought we needed to do before.

Enough to run the DRM... by Captain+Zep · 2006-02-20 02:39 · Score: 4, Insightful

Unfortunately most of the extra processing speed this gets you will be sucked up the all the DRM software running sefl-checks on itself, calling the mothership, and triple checking that you are licensed to excecute the next instruction.

So your computer will be nice and fast, just not any of your applications...

Z.

Re:Enough to run the DRM... by Macguyvok · 2006-02-20 03:40 · Score: 3, Insightful

Well, why that might be true for some, I've not yet seen any DRM software coming from the OSS camp. You all run DRM enabled AIM 6.6.6, I'll sit over here nice an happy running my gAIM 7.0 on my 23 teraherts AMD Zues 5400k+ with my 1.2 jigawatt powersuply. It'll run nice and fast.

Oh, and that's not to mention linux not having DRM. And before you tell me that I won't be able to play my DVD's, or mp3's, or whatever, I'll point out OggVorbis for audio files (no DRM in that, nor will there be) and I'll also point out the simple fact that 8 out of 10 hackers run linux. How could they live without their StarWars DVD's, or Doom3: The Rock's Back, Again.

Silly Consumer, DRM's for Little Girly Men.

--Macguyvok

--
--Mac "Nine point eight meters per second squared: The Best Damn Windows Accelerator, Ever."
Re:Enough to run the DRM... by Anonymous Coward · 2006-02-20 14:01 · Score: 0

Certainly Intel are already planning to use CPU real estate for their hardware DRM... so that you don't need to rely on motherboard designers, It'll be right there for you when you buy Intel (tm).
Re:Enough to run the DRM... by cfuse · 2006-02-21 09:39 · Score: 1

Unfortunately most of the extra processing speed this gets you will be sucked up the all the DRM software running sefl-checks on itself, calling the mothership, and triple checking that you are licensed to excecute the next instruction.
So your computer will be nice and fast, just not any of your applications...
DRM? More likely all my fresh new cycles will be eaten by running the Aero UI in Vista. My C64 seemed a lot faster than my Amd does today.
Cfuse's law: Any gains from Moore's law will be offset by bloat in Microsoft's latest OS.

New Metric? by BlueShirt · 2006-02-20 02:39 · Score: 1

From the article: Such a chip could store roughly 2,000 songs based on today's storage standards. Is this the new metric for machine capacity? Whatever happened to the number of books that could be held in RAM?

Re:New Metric? by shawn(at)fsu · 2006-02-20 02:44 · Score: 1

Back in my day size was measured in Libraries of Congress and we liked it that way.

--
500 dollar reward for tip(s) leading to the arrest of the person(s) who stole my sig.

Re:What's the minimum then? by PoconoPCDoctor · 2006-02-20 02:44 · Score: 5, Interesting

While the smallest chunk of silicon we could lay down would be one atom of it, there are things far smaller. In fact you can go something like 26 more levels of magnitude smaller before you start reaching the feasable limit of measurable existance. And yes, subatomic particles could theoretically be used in processors.

The process designation refers to the the distance between the source and drain in the FETs (transistors) on a processor. Keep in mind that this distance is by no means the smallest thing in the processor - the actual gate oxide layer is tiny by comparison, with Intel's 65nm process having only 1.2nm of the stuff. That's less than 11 atoms thick.

Found this on a thread at bit-tech.net forums.

--
"Let us raise a standard to which the wise and honest can repair" - George Washington

Yardsticks? I got yer yardstick. by Quiet_Desperation · 2006-02-20 02:44 · Score: 5, Funny

As capacity increases, new yardsticks are required. Eventually it'll go from number of songs to hours of porn, then hours of HD porn, then hours of full sensory VR porn experience, hours of holodeck recording and finally number of downloaded human personality matrices... of porn stars.

You can trust me on this. I have access to that interweb thing.

Re:Yardsticks? I got yer yardstick. by mr_Spook · 2006-02-20 03:03 · Score: 1

Porn or not, as bandwidth and hard drive capacity have increased over the years, people have found new ways to make heavy use of both. There was a time that spending a couple hours to download a text file wasn't unheard of. Now, full DVD rips are thrown about the net like it were nothing.

Despite the humor of the parent post, people always find ways to make hardware obsolete on their own.
Re:Yardsticks? I got yer yardstick. by khayman80 · 2006-02-20 09:54 · Score: 1

I have a feeling that I can already fit a huge "number of downloaded human personalities of porn stars" on my USB flash drive.

Moore is Less by dcw3 · 2006-02-20 02:45 · Score: 2, Insightful

I've heard the predictions for the end of Moore's Law, but haven't paid attention to the reasoning behind them. Is there some (sub)atomic barrier that is supposed to cause this? I was curious if further technological breakthroughs wouldn't prove these predictions incorrect. What would the predictions have been 20 or 30 years ago for our current tech? I doubt few, if any, were able to guess correctly.

--
Just another day in Paradise

Re:Moore is Less by PipeIsArt · 2006-02-20 03:57 · Score: 1

Of course, Moore's Law could eventually either prove or disprove String Theory. Taking us out a century or so, once the size of a microprocessor reaches a Planck length either Moore's Law or String Theory will fail.

--
I find that although many people are liberal in beliefs, they are conservative in actions.
Re:Moore is Less by Surt · 2006-02-20 04:47 · Score: 1

http://www.hal-pc.org/journal/03feb/column/baby/ba by.html

To summarize the portion of that article of interest: A silicon atom is .3 nm across. We are currently building transistor devices on 45nm processes. So if we reduce the process size to a single atom (and that's being generous: how do we control a device composed of one atom?), we'd achieve 150x density, in two directions, which would be 22,500 times improvement. That's enough for less than 15 more doublings, but I'll be generouse and give you the full 15. So if Moore's law is 18 months (and heck, I'll give you 24 months for doubling these days as things slow down, and remember when it was 12 months?) then we have 30 years left in Moore's law, before we hit 'devices' that are somehow magically made out of single atoms, yet still do the work we expect them to do.

It's relatively clear now that soon we'll have to learn how to build multilayer (3D) chips to keep making meaningful advances in conventional computer performance (there's always the possibility that quantum computers will make conventional computing performance meaningless).

--
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
Re:Moore is Less by Phase+Shifter · 2006-02-20 08:26 · Score: 1

To summarize the portion of that article of interest: A silicon atom is .3 nm across. We are currently building transistor devices on 45nm processes. So if we reduce the process size to a single atom (and that's being generous: how do we control a device composed of one atom?), we'd achieve 150x density, in two directions, which would be 22,500 times improvement. That's enough for less than 15 more doublings, but I'll be generouse and give you the full 15. So if Moore's law is 18 months (and heck, I'll give you 24 months for doubling these days as things slow down, and remember when it was 12 months?) then we have 30 years left in Moore's law, before we hit 'devices' that are somehow magically made out of single atoms, yet still do the work we expect them to do.
There are a few other points to consider.
1)Chips are almost pure silicon, but they require dopants to function. When you shrink the transitor size down to a cluster of a few hundred atoms or so, expect the properties to change, since you'll have to effectively make the impurities 1000 times more concentrated to maintain at leat 1 dopant atom per device. Shrink the transistor down to a half-dozen atoms, and the dopant is now a major component of the transitor (17%) instead of a trace impurity (0.001%).
2)The hole/electron associated with a dopant atom isn't tied down to a precise location in the crystal...It tries to stay in the general vicinity of the dopant atom, but due to the laws of quantum mechanics (and thermodynamics) these charge carriers tend to wander around the crystal a bit. If you attempt to pack devices too close together, this will result in charge carriers that may not even be in the same component as the dopant atom. This means the components won't function.
We'll have to abandon silicon before shrinking transistors this small, so we'll have to swich to some other strategy for increasing component density--3D component layouts, organic transitors, etc.

This isn't good news at all by B3ryllium · 2006-02-20 02:45 · Score: 3, Insightful

All this means is that AMD and Intel have to license the technology from a competitor. That's hardly good news for them, and it probably means higher CPU prices for us.

This isn't good news at all.

Re:This isn't good news at all by Anonymous Coward · 2006-02-20 03:01 · Score: 0

That's not the way the patent cabal works. The members of the oligopoly are already cross-licensed with each other. The goal is not to make big companies pointlessly write checks to back and forth to each other, but instead it's to keep any potential new players out of the game.
Re:This isn't good news at all by the+eric+conspiracy · 2006-02-20 04:18 · Score: 1

AMD has been licensing technologies from IBM for many years, including SOI etc. It hasn't seemed to hurt them much. Sometimes leveraging R&D across multiple manufacturers is a bvery good thing because if makes more money available for research.
Re:This isn't good news at all by Anonymous Coward · 2006-02-20 05:22 · Score: 0

What?! Technology improves and... it's... bad news???

Get the fuck back on the farm.
Re:This isn't good news at all by darkmeridian · 2006-02-20 06:59 · Score: 1

AMD and IBM already cross-license their IP. Silicon on Insulator was an IBM technology, remember, that AMD now uses.

--
A NYC lawyer blogs. http://www.chuangblog.com/

Well, NO. by Ancient_Hacker · 2006-02-20 02:49 · Score: 2, Interesting

Just being able to make thinner lines is not that huge a deal.

There's several large cans of whup-ass that have to be overcome before you can make IC's that much smaller:

Lines are 2-D thingies, but conductors are 3-D. Your etching technology has to get X times better to keep up with the line-drawing technology.
Same thing with the active components. If you try making the transistor half the old linear dimensions, you have 1/8th the volume of active silicon. This leads to all kinds of problems with leakage and power handling capability.
A line that's half as wide and half as thick has four times the resistance per unit length, and 1/4 the current-carrying capacity. You can try using a better conductor, but once you get to using copper, you're done.

And the programmers will just soak up all your extra speed by turning up the "ooooh" factor (See: Vista).

Re:Well, NO. by ChrisMaple · 2006-02-20 03:19 · Score: 2, Insightful

Current carrying capacity is important mostly for supply rails. In high complexity digital chips, the supply current mostly is routed on the highest metal layers, which are thicker than the layers near the transistors. These high layers are often almost completely dedicated to power distribution, so the lines can be quite wide.

--
Contribute to civilization: ari.aynrand.org/donate
Re:Well, NO. by lbrandy · 2006-02-20 03:31 · Score: 4, Insightful

* Lines are 2-D thingies, but conductors are 3-D. Your etching technology has to get X times better to keep up with the line-drawing technology.
* Same thing with the active components. If you try making the transistor half the old linear dimensions, you have 1/8th the volume of active silicon. This leads to all kinds of problems with leakage and power handling capability.
* A line that's half as wide and half as thick has four times the resistance per unit length, and 1/4 the current-carrying capacity. You can try using a better conductor, but once you get to using copper, you're done.

Why do I get the feeling that you actually have no idea what you are talking about, and neither do the people who modded you up. Etching, depositing, and lithography all go hand in hand when talking about an Xnm "process", therefore your comment about "thinner lines", in fact, makes no tangible sense. Lithography is the most difficult to shrink, not etching, so I'm really failing to see your point. It has been the main technical hurdle for the past 10 years.

Furthermore, the "conductors" in a processor aren't nearly as dependant on size as the silicon-feature construction. You can have an extremely layered chip with larger conductors if need be (and modern chips are), so both comment #1 and #3 are reasonably meaningless.

As for comment #2, yes, you are right: the "smaller transistor" problem is very well understood and it's the reason it takes so long to construct smaller and smaller processes, because the physics and effects must be taken into account. Not all transistors on a chip are the same size, nor can all transistors be shrunk. There is a reason that Intel doesn't slap it's PentiumIV plans into the new 30nm machine, and out comes a new chip. They have to go through and make sure that all the transistors that can be shrunk are, and none of those that cannot, are not. This is a reasonably non-trivial task, but it is not impossible, nor a "large can of whup-ass".

(PS: Thanks for the math lesson about 2d vs 3d in part 1. You might want to recheck part 3, with that in mind.)
Re:Well, NO. by phsdv · 2006-02-20 07:04 · Score: 2, Insightful

You are correct about the power lines, but not all power routing is done in thicker metal layers.
Besides do not forget that you need a lot of current to charge a very small capacitor very fast! Modern minimal sized transistors can switching a 10 to 100mA each. These currents have to go through a almost 100nm with (copper)line. That are still high current densities!
Going back on topic, the real amazing thing is that they can make very small lines (30nm) with light of a much larger wave length! Currently the industry can make 65nm lines using light with 193nm wavelenght! Think about it, this is line with a size of 1/3 of the wave length used! IBM has probably used 157nm wavelength to make 30nm lines, which means an factor of 5!
Using much smaller wavelengths is a problem there the light (if you can still call it that) will be absorbed instead of transmitted by conventional lenses. See for example http://www.llnl.gov/str/Sween.html for more info.
Thus special (expensive) lenses are needed when using Extreme-UV. The real news in this article is that it they made it possible with 'conventional' lenses and light sources!
Re:Well, NO. by Ancient_Hacker · 2006-02-20 08:20 · Score: 1

What I was trying to convey was the that the simplistic original article didnt even begin to convey the scope and depth of the challenges. Just being able to draw narrower lines isn't the be-all and end-all. If not several cans of whup-ass, at least one of worms.
Re:Well, NO. by Just+Some+Guy · 2006-02-20 09:15 · Score: 1

You can try using a better conductor, but once you get to using copper, you're done.
You misspelled "silver", and ignored whole classes of relatively exotic materials like carbon nanotubes.

--
Dewey, what part of this looks like authorities should be involved?
Re:Well, NO. by Ancient_Hacker · 2006-02-20 10:23 · Score: 1

Er, No.
The difference in conductivity between copper and silver is miniscule. 1.58 versus 1.68. That's 1 part in 16. According to Moore's "Law", that gives you a month and six days of breathing room.
And it's heartwarming to see such undiluted faith in nanotubes. Some might have doubts, as nobody's yet intentionally made even a millimeter of ballistic nanotube at any price, whilst a CPU chip would need quite a stretch of them, all perfect, for under $5.

Re:Moore's Law is so 1998 by Anonymous Coward · 2006-02-20 02:50 · Score: 0

it's not about speed, but complexity - from Wikipedia:

Moore's law is about the empirical observation that at our rate of technological development, the complexity of an integrated circuit, with respect to minimum component cost, will double in about 18 months.

What happens... by sphealey · 2006-02-20 02:50 · Score: 3, Funny

What happens when they get to -1 nm then? Can they keep going smaller?

sPh

Re:What happens... by publicworker · 2006-02-20 03:13 · Score: 1

What happens when they get to -1 nm then?

Imaginary chips!
Re:What happens... by PFI_Optix · 2006-02-20 03:13 · Score: 2, Funny

Yes. Intel will release the new Singularity chip some time around 2016.

AMD fanboys will promptly inform everyone that it sucks.

--
120 characters for a sig? That's bloody useless.
Re:What happens... by Anonymous Coward · 2006-02-20 10:15 · Score: 0

Depends if they used 'signed' or 'unsigned', and the word size.

According to my calculations, we'll have 4.29m chips within the decade.

Re:Moore's Law is so 1998 by Scarblac · 2006-02-20 02:51 · Score: 4, Informative

Moore's law is about the number of transistors on a chip. It states nothing at all about speed.

--
I believe posters are recognized by their sig. So I made one.

Moore's law is not about speed by ebcdic · 2006-02-20 02:54 · Score: 1

It refers to the "complexity" of integrated circuits, not their speed. Complexity is taken to refer to the number of components, which corresponds to their size. Recently it has become more difficult to translate smaller components to greater speed, but that does not contradict Moore's law.

Re:What's the minimum then? by edxwelch · 2006-02-20 02:56 · Score: 1

Well, if the gate layer is the smallest thing in the transistor and it is 11 atoms wide and 1 atom is the smallest measure, then smallest transistor theortically possible is 65nm/11 = 6nm

Re:What's the minimum then? by bronney · 2006-02-20 02:56 · Score: 1

Thanks. So you're saying that our first challenge is to have the "distance between the source and drain in the FETs" = 1 atom thick. Then the next challenge would be developing a whole new process in using subatomic particles to switch. That of course, we're still using this "transistor" method when we get there.

Re:Moore's Law is so 1998 by LLuthor · 2006-02-20 02:58 · Score: 1

Moore's law says nothing about speed or frequency of the CPUs... merely that the transistor count will double every 18 months. And, so far it has. My Athlon-64 X2 4800+ has double the transistor of my best CPU 18 months ago (give or take a bit for cache).

--
LL

"Moore's law" is not a law of nature by MZ80K · 2006-02-20 02:59 · Score: 2

"Moore's law" is not a law of nature, it is a marketing strategy of Intel.

Re:"Moore's law" is not a law of nature by NutscrapeSucks · 2006-02-20 08:33 · Score: 1

You sir, were moderated unjustly. If you read Moore's original paper, he's speaking directly about marketing.

(And while Moore's Law is phrased as an engineering challenge, Intel has historically used it as a form of "planned obsolecence" to drive demand for new CPUs.)

--
Whenever I hear the word 'Innovation', I reach for my pistol.

Re:What's the minimum then? by lbrandy · 2006-02-20 03:00 · Score: 4, Informative

Well, if the gate layer is the smallest thing in the transistor and it is 11 atoms wide and 1 atom is the smallest measure, then smallest transistor theortically possible is 65nm/11 = 6nm

You are confusing dimensions. When intel refers to 65nm processes, they are talking about length and width ability to carve out features. Oxide layers "thickness" operates in the third dimension ("height"?) to provide resitant layers. It is much smaller then 65nm. Actual atoms are about 200 picometers in "width".

Moore of this kind of thing! by Toreo+asesino · 2006-02-20 03:01 · Score: 1

/coat

--
throw new NoSignatureException();

Re:Moore of this kind of thing! by ScrewMaster · 2006-02-20 03:44 · Score: 1

Yes, they'll be building CPUs out of a new subatomic particle called a "Mooreon."

--
The higher the technology, the sharper that two-edged sword.

MOD PARENT UP! by engagebot · 2006-02-20 03:01 · Score: 0, Flamebait

Best points made so far. MOD PARENT UP!

--
Han shot first.

Moore's law died years ago. by Sheepdot · 2006-02-20 03:07 · Score: 4, Informative

I think it was in 2000 that a /. patron actually listed the "complexity"-related proof that Moore's law died in 2000, but here's my contribution:

Who said what?
California Institute of Technology Professor Carver Mead was the one who dubbed it Moore's Law, a lofty title Moore said he was too embarrassed to utter himself for about 20 years. David House, a former Intel executive, extrapolated that the doubling of transistors doubles performance every 18 months. Actually, performance doubles more like every 20 months. Moore emphatically says he never said 18 months for anything.

The rule also doesn't apply to hard-drive densities or to the growth of other devices. "Moore's Law has come to be applied to anything that changes exponentially, and I am happy to take credit for it," Moore joked.

From:
http://news.com.com/FAQ+Forty+years+of+Moores+Law+ -+page+2/2100-1006_3-5647824-2.html?tag=st.num

This is not about mhz ratings, though for a while these were doubling along the same rate as transistors per square inch were. Moore's comments were about integrated circuit "complexity" minimum component costs, which, if you are talking about transistors, has remained reasonable accurate. If you are talking about mhz per dollar, then you're going to find this is not accurate at all.

Long story short, if you had a 2 ghz machine in early 2003 and you're wondering why you aren't on an 8 ghz machine now, it's because mhz ratings have NOTHING to do with Moore's Law. Which is why I suggest referring to the Wiki entry on it.

Also important is Kryder's Law for HD storage capacity. Within a decade or two we may be able to store all creative works ever created on one drive.

Case in point: Hard drives increase a thousand-fold in storage space every 10.5 years. In 1996 I purchased a Compaq computer with a 1 gig drive. That was an insane amount of space at the time, but now, 10 years later, it looks like I may be able to purchase my first TB drive soon.

Re:Moore's law died years ago. by marcosdumay · 2006-02-20 05:12 · Score: 1

"Also important is Kryder's Law for HD storage capacity."

I followed your link. Now I understand why I never heard about this law... That graph is anything but a straight line. When you have a cluster of points, a gap, and another cluster of points, the easiest thing to come out with is a straight line. On this case, you have the two culsters seppareted by a gap, and not even this way it look straight.

But thanks. Now I know that HD sizes don't folow an exponential law.

--
Rethinking email
Re:Moore's law died years ago. by Anonymous Coward · 2006-02-20 05:42 · Score: 0

Within a decade or two we may be able to store all creative works ever created on one drive.
Not once I get my pr0n collection upgraded to HD!
Re:Moore's law died years ago. by nasch · 2006-02-20 06:13 · Score: 1

Are you saying you think someone just eyeballed the points and drew in a straight line? Don't you think it's more likely that a charting program calculated a best-fit line? We are of course missing information indicating how much the data deviate from the line, but still... I'm missing your point as well.
Re:Moore's law died years ago. by marcosdumay · 2006-02-20 10:01 · Score: 1

Of course the line may be calculated by a best fit algorithm. I never questioned that. The problem is that the points dont fit well above a straight line. So they should be described by something else.

My other point is that when you lack the middle points (there is a gap there), you can't reliably model it. It will probably look like anything that you want*. And even so, the author had the bad luck of getting points that don't look like a straight line.

* That is not true if you have precise measurements, but it is not the case.

--
Rethinking email
Re:Moore's law died years ago. by nasch · 2006-02-20 11:52 · Score: 1

So you're saying that 1) the line may be a best fit, but the fit isn't very good, and 2) there's a gap, and 3) the points don't model a straight line. Is that correct? If so, how do you judge 1? By eyeballing? I didn't see any data about standard deviation or the like, so how do you know how good the fit is? Can't argue with number 2, could be a problem. And for number 3, are you saying the data aren't linear? If so, that's totally consistent with the chart - it also indicates the trend is not linear.
On the other hand, if I misunderstood you, please set me straight.
Re:Moore's law died years ago. by marcosdumay · 2006-02-20 12:45 · Score: 1

1: Yes, by eyeballing. It is not enought prove that the data fit the model. But it can be enoght to perceive that it doesn't fit. He arguees that there is a line there, but everything I can see is some kind of curve. I'd like to see the correlation of the fit, so I could numericaly prove my point, but it is really not needed.

3: That came from item 1. I'm saying that it is not an exponential law, because the graph is not an straight line.

--
Rethinking email

Re:Punctuated Equilibrium - Phase Transition in Mo by dslauson · 2006-02-20 03:13 · Score: 1

If you are simply talking about Moore's Law in terms of processing power, there are other places to gain improvements rather than just compactness of chips. There is also parallel processing technology, which is still steadily improving.

Then, far off over the horizon, there's the possibility of quantum computing, which would make for a rediculously huge surge in processing power all at once.

That's fundamentally how Moore's Law works: as soon as the current paradigm starts to get maxed out, we simply shift to another paradigm.

Halt! Zeno Police. by Anonymous Coward · 2006-02-20 03:14 · Score: 0

I'm sorry, you've exceeded the Zeno's Paradox limit.

Please go back half way.

Re:What's the minimum then? by zenwrench · 2006-02-20 03:18 · Score: 0, Offtopic

According to wikipedia ... bill gates never actually said that: wiki

but me still like

Materials science by sphealey · 2006-02-20 03:19 · Score: 1

I am guessing they will use resublimated Thiotimoline as the base material, proving once again that Asimov was ahead of his time.

sPh

Re:Moore's Law is so 1998 by Anonymous Coward · 2006-02-20 03:20 · Score: 0

The parent does have a point though. Even if Moore's law holds and the transistor count doubles every 18 months, the law is useless unless we can get proportional utility from the transistors. Otherwise we just end up with a chip twice as complex (and more costly) than its just as useful younger, simpler cousin.

Re:Punctuated Equilibrium - Phase Transition in Mo by QuantumFTL · 2006-02-20 03:31 · Score: 4, Interesting

If you are simply talking about Moore's Law in terms of processing power, there are other places to gain improvements rather than just compactness of chips. There is also parallel processing technology, which is still steadily improving.

There are many important algorithmic problems that are inherently serial. Some things are mathematically impossible to parallelize. Also limitations caused by enforcing cache coherency, communications interconnects, and resource access synchronization/serialization create bottlenecks in parallel systems. The astrophysics simulation code that I paralellized is almost entirely math operations on large arrays (PDE solving), however there are diminishing returns past 48 processors due to communications latency. Better programming techniques can push the limit of this, however it is difficult to design software that mitigates the effects of this kind of latency without many man-hours spent to handle it.

Then, far off over the horizon, there's the possibility of quantum computing, which would make for a rediculously huge surge in processing power all at once.

I mentioned this in my post, however there is a bit of a catch. Quantum computing, practically speaking, is only useful for certain problems - problems that are "embarassingly parallel." QC does not help with fundamentally serial problems, and is likely to be impractical beyond a critical number of qubits, due to quantum incoherency, even quantum error correction can only stretch so far. Great for cryptography/number theoretic operations, and probably many optimization problems (scheduling perhaps?) but certainly not for standard computation. Problems (like database queries) that require large amounts of data to be stored in a quantum coherent fashion are unlikely to be practical.

"That's fundamentally how Moore's Law works: as soon as the current paradigm starts to get maxed out, we simply shift to another paradigm."

Ahh, but that's just it - there is a cost to the switch in terms of both time and money. What I am saying is that yes, we can continue to change paradigms whenever we hit a limit, however these transitions will be very expensive and will cause "delays" during which little improvment on shipping computer technology will be seen.

Re:Yardsticks? I got yer yarkstick. by hachete · 2006-02-20 03:33 · Score: 1

Ah, the true meaning of teh Yarkstick

--
Patriotism is a virtue of the vicious

dMoore/dt by Doc+Ruby · 2006-02-20 03:37 · Score: 1

I wonder whether we'd be getting faster productivity increases if Gordon Moore had included an extra term in his observational Law to accommodate acceleration of the rate, not just the rate, of transistor density increases. Moore's Law is partly self-fulfilling (as an upper limit), because it's the highest consensus rate expectation, especially among managers who control budgets.

--

--
make install -not war

Re:What's the minimum then? by ZonkerWilliam · 2006-02-20 03:43 · Score: 1

Can you be more precise than "11 atoms thick". This erronous since each type of atom ie. Hydrogen, Carbon, silicon, are different sizes.

Great, now let's get software on board by ZachPruckowski · 2006-02-20 03:45 · Score: 2, Insightful

Now the problem here is that software seems to be getting less efficient. Even with faster processors, checking your email, web browsing and word processing now takes a lot more RAM than it used to. If software was getting more efficient, or at least holding to the same level, we'd be a lot farther ahead now.

Re:Great, now let's get software on board by eyepeepackets · 2006-02-20 04:28 · Score: 1

That depends on the software you choose to use. I'm not seeing any bloating problems on the software side (2.4 kernel, Slackware, WindowMaker, Opera.)

If you're using commercial software from Redmond, then yes, that particular corpse is bloated like a five day old dead Australian.

Disclaimer: I choose to use an Australian because they tend to have a good sense of humor, there's no politically correct downside to choosing them, and they make good belts.

Cheers.

--
Everything in the Universe sucks: It's the law!
Re:Great, now let's get software on board by ZachPruckowski · 2006-02-20 04:32 · Score: 1

I was talking about products in mass use. I personally am 100% Redmond free, but like 95% of the computer using world isn't.
Re:Great, now let's get software on board by cyclomedia · 2006-02-20 04:42 · Score: 1

true, but then again i'm still runing windows 98 and it zips along very bloody nicely on a 2.8ghz p4 i can tell you.

--
If you don't risk failure you don't risk success.
Re:Great, now let's get software on board by joss · 2006-02-20 05:10 · Score: 1

> they make good belts.

Really ? I know crocadiles make good belts, but I didnt know you could, er, cut out the middle man.

--
http://rareformnewmedia.com/
Re:Great, now let's get software on board by Just+Some+Guy · 2006-02-20 10:13 · Score: 1

Now the problem here is that software seems to be getting less efficient.
Really? Many major software releases (OS X, KDE, etc.) seem to be getting more efficient with each release. Now, if you want to argue that software is doing a lot more than it used to, then I'd have to agree.
The applications I used in the '90s were "dumb" enough to only read/write to the local hard drive. I didn't possess a program that had strong cryptography. No one used array bounds checking because it was too expensive. Video games wrote directly to display memory. Using 44.1kHz audio samples for system notifications was a fantasy. I had to defrag my hard drive. And so on, and so forth.
Compare to today:
My text editor knows WebDAV and sftp (via shared system components). My email program uses TLS. Programs are increasingly written in "safe" languages. Hardware is abstracted. High-quality media is pervasive. My filesystem is based on B* trees.
Each of those things incurs a small overhead, but I would never go back to the "good old days" (hah!). Rather, my position is that hardware is finally catching up with the future we'd always planned but couldn't afford to implement.
Besides, the only people who still think yesterday's systems were more responsive or efficient are the ones who haven't revisited them. I use to be an Amiga junkie - until I made the mistake of firing up an emulator that outspecs and outbenches my last real Amiga in every metric imaginable. Good grief, how that sucked (although it was still better than any PC :-P ).

--
Dewey, what part of this looks like authorities should be involved?

Why small? by briglass · 2006-02-20 04:00 · Score: 2, Informative

While this question will undoubtedly reveal my limited understanding of computer engineering, I will ask it anyway... Why is the industry obsessed with getting smaller chips? There's plenty of room on my desktop for a hefty five-inch or even ten-inch diameter chip if it meant greater processing power and/or speed. Is the reason that they shoot for smaller chips that by making the chip smaller and smaller, it can run more calculations per second just in virtue of the speed of the electrons through the circuitry? Even so, I hear about people joining processors together to increase speed/power... so why not shoot for utilizing older technology to create larger yet better chips?

--

----
"Those who quote others are more likely to one day be quoted" -Tom Planter

Re:Why small? by corngrower · 2006-02-20 04:17 · Score: 1

Quick answer: Once the size of a chip gets to a particular size, it's more economical to split the functionality among several chips. This is because defects exist in the silicon crystals of the semiconductor wafer. As the size of a chip increases, there is an expodentially increasing chance that the chip will have one of these defects in it, ruining the chip. With Small chip sizes, most of chips on a wafer are good. With large chip sizes most of the chips on a wafer ar bad. About 1cm^2 seems to be the practical upper limit on chip sizes these days.
Re:Why small? by Surt · 2006-02-20 05:00 · Score: 1

There are a number of real reasons you don't want to go this direction as a chip maker:

1) The chance of a chip killing flaw is proportional to the area of the chip.
2) The cost of manufacturing each chip is proportional to the area of the chip.
3) The time to transmit a signal across a chip is proportional to the length of the chip. We can build multi-chip and multi-core setups, but they are slower due to between CPU communication overheads.
4) Clock distribution is a challenge over a large chip (this may possibly be solved by clockless chip designs in the future).

I'm sure there are more, those are just the ones off the top of my head.

--
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
Re:Why small? by necro81 · 2006-02-20 05:24 · Score: 4, Interesting

There are several reasons why the industry is focused on smaller. I do not work for a semiconductor manufacturer, so some of my information may be a little off.

1) Defects and Yield. Most processors are manufactuered out of silicon wafers 300 mm in diameter. The wafer is very pure silicon (before they start doping it), and the crystal structure is one of the most perfect and regular that humankind has ever been able to produce (at least on a large scale). The industry doesn't do this merely to be perfectionist - it costs a LOT of money and infrastructure to do it - but simply because defects in the crystal structure and silicon purity result in a non-functional chips. The statistics and probabilities behind how many defects get scattered on a wafer, and how many potentially useful chips do those defects knock out has been heavily studied by the industry. The yield that one gets from a single wafer that has many chips on it is a function of defect density and chip size (and other things). A larger chip naturally has a greater chance of having a defect than a smaller chip. There isn't much more that the industry can do to reduce the number of defects on a wafer. In order to increase yield, one of the things the industry banks on is decreasing the chip size. The yield for, say, op-amps (which are very tiny chips) is much higher than for full-blown processors.

2) Signal Distance. The upper limit of speed for an electronic signal in a chip is the speed of light. That's really fast, but not infinite. In fact, compared to the clock speed of the chip itself, the speed of light becomes significant. The speed of light in a vaccum is 3 * 10^8 m/s. In one nanosecond, light travels 30 cm. For a 4 GHz processor, light can travel only 7.5 cm between clock cycles. In truth, the electronic signals in the chip travel slower than that. So, the distance between various parts of the chip become significant. For a chip as large as several inches, it can take quite a long time, many clock cycles, for bits to make it from one end to the other. Wasted clock cycles = reduced performance. So, in order to continue increasing performance, the industry has worked very hard to keep the size of processor chip very small, so that it takes very little time for signals to travel across it.

3) Power. It would take a while to explain the physical reasons behind it (see an VLSI or semiconductor textbook for a full analysis), but the operating voltage of a transistor goes down as its physical size goes down. It used to be that 5 V was the working voltage of most all transistors. Then it moved to 3.3 V. Nowadays, the core voltage of most processors is around 1 V. As the operating voltage has decreased, so too has the power dissipation per transistor. The deceasing feature size of transistors and photolithographic techniques is largely to thank for this. The reason that processors now dissipate such a large amount of heat is that, even though the per transistor power has decreased, the number of transistors in the chip has increased more rapidly. If one tried to make a P4 chip using 350 nm techniques (which used to be the standard feature size les than a decade ago), the chip probably would dissipate many hundreds of Watts.

4) Speed. One would again have to check out a VLSI textbook for a full explanation, but (physically) smaller transistors can switch states faster than large ones. While clock speed is far from the be-all, end-all measure of processor performance, it is generally true that faster transistors result in faster performance (hence the whole notion of overclocking). Using the szame "P4 made using 350 nm technology" example, it would be impossible to run such a chip at anything close to 4 GHz. In fact, I doubt you'd be able to get it to run at even 1 GHz - the transistors would simply be too slow. I don't recall exactly when 350 nm was the standard technology used by the industry, but I imagine that you'd find it coincided roughly to the times when chip speeds were mea
Re:Why small? by Anonymous Coward · 2006-02-20 09:02 · Score: 0

Why is the industry obsessed with getting smaller chips?
Money.
If you have a wafer which is 300mm in diameter, you have ~70,000mm to make chippies out of (inefficiencies around the edges, &c.; I'm rounding down). If each chip is 25x25mm, that's 113 chips/wafer. At $250/chip, that's a bit over $28,000/wafer (gross). make the chips each 50x50mm, and you're either going to have to settle for a bit over $7,000/wafer, or raise your price to $1,000/chip.

Intel at 65nm, AMD in 07, a year and a half late. by Anonymous Coward · 2006-02-20 04:00 · Score: 0

AMD made a borked business decision and so they are going to be stagnate for 2006, loosing all their gains as Intel catches up, then passes them. Only just starting the process for even lining up the loan for buying the equipment to make 65nm parts. Woof. I did find it odd the mexican guy they had running AMD stepped down so soon. They'll be treading water all year from this screwup.

Moore's law is universal, not empirical by koroviev+(begemot) · 2006-02-20 04:02 · Score: 2, Interesting

Actually the more interesting thing about Moore's law (in terms of total processing power) is that it holds way further back than most people think. Mechanical calulator's total numbers and performance (like Charles Babage's difference engine) were also in accordance with Moore's law, and the two curves fit together quite nicely with the advent of the "many women" approach to computing and electronical computers. Even clock-making reflects Moore's law in the last hundreds of years - in terms of unit numbers, clock sizes, element sizes (the size of the gears), the switch to electronic watches, etc..

Re:Moore's law is universal, not empirical by wildsurf · 2006-02-20 09:33 · Score: 1

with the advent of the "many women" approach to computing

Boy, things have changed.

--
Weeks of coding saves hours of planning.

Re:What's the minimum then? by corngrower · 2006-02-20 04:05 · Score: 1

You're not familiar with how they make CMOS semiconductors then. The gate dielectric is almost always silcon dioxide. So consider a layer of silicon dioxide crystal 11 molecular layers thick. That means if you make a processing mistake and your layer is 12 molecular layers thick, you're going to notice it in terms of circuit performance.

We call it Moore's law... by Kjella · 2006-02-20 04:11 · Score: 2, Insightful

...for the same reasons we call it Murphy's law. The world would be a pretty terrible place if absolutely everything that could ever possibly go wrong, did. In both cases it's just a perception that things behave in a law-like manner even though there's obviously no scientific basis and with plenty of counterexamples. As far as technology predictions goes, it is disturbingly accurate, it follows a mathematical formula as most laws do... so we call it a law. It's a joke, laugh.

And the rub of it is exactly what you say - it seems to just keep going and going, despite its obvious unsustainability. My dad used an osciloscope on single bits in radio tubes, can you imagine what they said in the 60s? 70s? 80s? 90s? "This can't go on". Moore's law seemed (seems?) to stand above the laws of nature. That's what makes it so intriguing. But it has far more to do with social science than natural science...

--
Live today, because you never know what tomorrow brings

Re:What's the minimum then? by bbrack · 2006-02-20 04:13 · Score: 1

The process designation refers to the the distance between the source and drain in the FETs (transistors) on a processor. Keep in mind that this distance is by no means the smallest thing in the processor - the actual gate oxide layer is tiny by comparison, with Intel's 65nm process having only 1.2nm of the stuff. That's less than 11 atoms thick.

Not quite correct - the process designator refers to the size of a minimum-width metal line, the physical gate length is usually significantly smaller

130nm chips that I have either worked on or seen have gate widths ranging from 50nm - 90nm
At 90nm, this has shrunk to 35-60nm or so

Re:What's the minimum then? by Anonymous Coward · 2006-02-20 04:17 · Score: 0

Atoms don't actually vary in size all that much. From the lightest to the heaviest, the diameter difference (which isn't easily defined in any case) is no more than a factor of 2. The reason is the more highly-confined electronic states as the nuclear charge increases.

The elements found in silicon-based devices are all of a very similar diameter.

oh its a law huh? by Anonymous Coward · 2006-02-20 04:20 · Score: 0

Why is it always called Moore's Law? Shouldnt it be Moore's Theory, or Morre's Observation?

Re:oh its a law huh? by Spy+der+Mann · 2006-02-20 06:52 · Score: 2, Funny

Shouldnt it be Moore's Theory, or Moore's Observation?

Yeah, Moore or less :)

Off topic by Anonymous Coward · 2006-02-20 04:22 · Score: 0

Best Buy here in the US is running a commercial on TV and radio advertizing upgrades for computers including Seagates 500Gb hard drive. In the radio add a woman states how she has become a deleter because of not enough harddrive space. The radio commercial states that she wakes up at night wondering what she can delete next.

The announcer goes on to state that Seagates 500Gb hard drive can hold up to 100,000 digital photos.

Yes, I typed that right, the ad states the 500Gb drive can hold up to 100,000.

Those must be some really fricking huge digital photos.

Has anyone else heard or seen that commerical in other parts of the country. I'm in Atlanta Ga.

Re:Off topic by Anonymous Coward · 2006-02-20 04:33 · Score: 0

I have heard it in the morning weekdays ~7am CST outside Chicago. Possible stations: 104.3, 95.9, 100.7, 97.1, 97.9, 94.3 (IIRC, some digits may be off). It sounds just as stupid here. NO, I will NOT GET satellite radio (to whomever was thinking that -you know who you are).
Re:Off topic by databyss · 2006-02-20 04:42 · Score: 1

What?!? That's only like 512MB per photo!

You make it sound like everybody doesn't take snapshots with their Hubble.

--
Hmmm witty sig or funny sig? Maybe elitest techy sig!
Re:Off topic by swillden · 2006-02-20 06:33 · Score: 1

The announcer goes on to state that Seagates 500Gb hard drive can hold up to 100,000 digital photos. Those must be some really fricking huge digital photos.
5MB? That's not that large for a mid to high-end digital camera. It's a little low if you shoot RAW. If they assumed less than 2MB per photo (250K photos), I'd consider it false advertising.

--
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Re:Off topic by swillden · 2006-02-20 06:37 · Score: 1

That's only like 512MB per photo!
Hehe. Yer off by two orders of magnitude. 100,000 x 500MB is 50TB, not 500GB.

--
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Re:Off topic by databyss · 2006-02-20 08:26 · Score: 1

Damn... I'm 0 for 2 today. I hereby decree that I am dumb for today.

5MB a pic doesn't sound too unreasonable then for very high quality images at large resolutions like a lot of cameras take today.

--
Hmmm witty sig or funny sig? Maybe elitest techy sig!
Re:Off topic by raygundan · 2006-02-21 03:20 · Score: 1

Agreed-- my not-even-remotely-high-end digicam takes 2.5MB photos in jpeg and ~15MB photos in tiff. Seems like a fair estimate.

Re:What's the minimum then? by trentblase · 2006-02-20 04:36 · Score: 1

So untrue -- 65nm means a 65nm gate length. That means a 65nm wide poly (gate width is usually much higher than gate length). Minumum M1 width may be as high as 100nm in this process.

or they can use x-ray lithography... by YesIAmAScript · 2006-02-20 04:38 · Score: 1

This is only one way to make 40nm chips.

Additionally, maybe they'll pull off a patent swap, or will make other refinements to the process and contribute them in exchange for a reduction (or elimination) of license fees.

Or maybe since JSR Micro is a supplier to fabs, if you buy the exotic quartz crystal lens and other equipment from them (maintenance contracts?), perhaps JSR Micro will give the patent license for the process for free.

Or maybe they won't patent it, or the processor making chips can be altered in so may subtle ways that it's easy to get around them.

And apparently by "they" in this document so far I mean Intel, because AMD doesn't fab their own processors. IBM typically fabs them for them, and it's already announced AMD has developed a 65nm process with IBM.

All in all, it's very possible that this just isn't unusual, that AMD (in their non-CPU fabrication) and Intel have had to pay various license fees or deal with certain patents and fab equipment/technology companies on every process in recent memory.

--
http://lkml.org/lkml/2005/8/20/95

Re:Punctuated Equilibrium - Phase Transition in Mo by the+eric+conspiracy · 2006-02-20 04:42 · Score: 1

Yes, there are limits as to how far things can be paralellized. However for most common uses of computers we are far from those limits - even many of the languages that are commonly used don't support threading as a standard feature, or the support is not robust. How many languages support loop parallelization as a standard optimization?

Progress is being made though - for example computing languages such as Java have been adding support for atomic variables and other faciliites that reduce or eliminate the need for serialization.

Fight to Regain the Buldge! by ratboy666 · 2006-02-20 04:44 · Score: 1

Thank you! Very insightful post.

And you can quote me.

Now, on to some observations. We have been at a state of equilibrium now for a few years.

It is slightly difficult to determine exactly what the bounds are, because we are in it right now. I am guessing that the "slowdown" started around the time of the Pentium Pro ('96?).

The "clue" was the introduction of "Beowulf" clusters where processing is balanced with communications overhead.

Intel is fighting this with Itanium, SUN with Niagra.

I suspect that the new "Moores Burst" will begin with death of the cluster as supercomputer. I have always predicated that Itanium (or something architecturally similar) will win (and the pun is intended).

Ratboy

--
Just another "Cubible(sic) Joe" 2 17 3061

Re:Fight to Regain the Buldge! by HiThere · 2006-02-20 07:46 · Score: 1

It doesn't work like that. When the new round of more efficient CPU chips arrive, the technology used to create Beowulf clusters will just be used to cluster THEM, vastly increasing the speeds. It will, however, limit the use of clusters, as single chip computers will necessarily be simpler, and so as Beowulf clusters move higher up the chain, they'll lose more of the low end.

There will always be a need to cluster computers in a high latency manner to deal with intractable problems...and only quantum computers could hope to deal with NP hard problems, if them. (As has been pointed out, they suffer many severe limitations as to what they are capable of that will likely always restrict them to niche applications...but they could make just dandy specialized ganglia, unless their environmental constraints are too severe.)

--

I think we've pushed this "anyone can grow up to be president" thing too far.
Re:Fight to Regain the Buldge! by ratboy666 · 2006-02-21 05:57 · Score: 1

As long as interconnect speed paces processor power at a particular ratio (or better), the "beowulf" loosely connected cluster remains viable as a supercomputer. If the ratio tips, it doesn't work.

As an example -- you generally wouldn't think to build a "beowulf cluster" using an interconnect of 9600 baud modems, right? Even though you could, the cluster would, in general, not gain much.

In '96 (or thereabouts), such a loose cluster composed of Pentium Pros, with 100MB interconnects could make some sense. Now, we have optical interconnects, and are generally 1 order of magnitude in performance beyond that.

My point was that what we are looking for is a processor leap that leaves the interconnect "in the dust", thus making the loose interconnect concept silly again.

We aren't there yet. "Quantum Computing" won't be it, either. What is needed is an attack on ILP (instruction level parallelism). Which forces architectural changes. And programming changes. "C" (and its ilk) won't cut it, because algorithms are not exposed at a high enough level. Which was the big failing of Itanium.

If processor speeds could be increased by an order of magnitude TODAY, "beowulf clusters" cease to be relevant.

Ratboy.

--
Just another "Cubible(sic) Joe" 2 17 3061
Re:Fight to Regain the Buldge! by HiThere · 2006-02-21 06:31 · Score: 1

Actually, you would. The precise interconnect protocol would change, but remember that ARPAnet was built in a world with slow connections. True, the amount of sharing changes with a change in relative connection speeds, but even with extremely slow connections, some problems can be parcelled out into pieces. To pick a stupid, but easily explained, example:
You want to find the best chess game.
You have n computers.
So what you do is you assign each computer one of the openings from MCO, and you assign another to search games that DON'T start with one of the openings. I think MCO runs to over a thousand openings now, tell each computer to find the best game that it can within x amount of time. At the roundevue select the best. Note that interprocessor communication is very limited in this model.

--

I think we've pushed this "anyone can grow up to be president" thing too far.

Re:Exponential Growth by vertinox · 2006-02-20 04:46 · Score: 1

If this so-called law were to continue unabated for a couple of centuries, the number of transistors in a chip would exceed the number of atoms on planet earth. Clearly, a limit is going to be reached well before that happens.

I'd can't seem to find the source right now (its burried someone on wiki about human over population) but...

If the rate of human births to that of the population remains the same, there will be more humans than (guestimated) number of atoms of the universe in about 17,000 years.

Now explain to me how there can be more humans than atoms in the universe and I'll explain to you how to divide infinity by zero. ;)

But seriously, Moore's 'Law' holds true because it is a self fufilling prophecy. However, this is a good thing, because better the computers and the 'moore' of them the more money I get paid in IT these days and the sooner I get virtual sex when I am 60 years from now and butt ugly.

--
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)

Re:Punctuated Equilibrium - Phase Transition in Mo by vertinox · 2006-02-20 04:49 · Score: 2, Insightful

Some things are mathematically impossible to parallelize. Also limitations caused by enforcing cache coherency, communications interconnects, and resource access synchronization/serialization create bottlenecks in parallel systems.

Explain the human mind, then.

--
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)

Moore's Law is aready dying by smilindog2000 · 2006-02-20 05:03 · Score: 1

The likes of Intel will be able to afford 32nm. The rest of us are mostly still using 180nm and larger, where the cheapest consumer transistors are currently fabbed in China. The rest of the electronics industry already not benefit much from 65nm, so the great engine of Moore's has already failed us.

--
Beer is proof that God loves us, and wants us to be happy.

Re:Moore's Law is aready dying by iggymanz · 2006-02-20 08:01 · Score: 1

no, Moore's orignal statement was about complexity of integrated circuit with respect to minimum component costs, and even our cell phones chips roughly follow the rule.
Re:Moore's Law is aready dying by smilindog2000 · 2006-02-20 09:39 · Score: 1

The phrase "minimum cost components" was meant to describe the cheapest way to build a system. If you look inside modern consumer products, mostly you will find 180nm and larger. The cheapest transistors you can buy today are in 180nm on $600 8-inch wafers made in China. If you try to buy wafers at any smaller geometry, 130nm or smaller, you pay more per transistor. I got a quote two years ago for 12-in .13u wafers made in the US: $10K/wafer. The larger wafers in the smaller geometry have 4x the number of total transistors, but cost 16 times as much. The math doesn't work out.

In many applications, we're willing to pay more. For example, a cell phone manufacturer recently told me he shrank to 90nm to save power. Also, space is at a premium in a high-end cell phone. However, these phones are not the cheap ones. The cheapest cell phones still rely on older process technology. I think the way Moore said it, he would have used the cheaper cell phones that still use 180nm technology.

Anyway, if all we care about is the Pentium, yes Moore's Law is still on track and probably will remain so for some time. However, the impact that Moore's Law use to have on the electronics industry as a whole has already faded dramatically.

--
Beer is proof that God loves us, and wants us to be happy.

What should we call it? by Beryllium+Sphere(tm) · 2006-02-20 05:05 · Score: 1

>It's not a law. It's just incorrectly called a law.

Writing with precision is good. Exponential growth of transistor counts is not a statute and it's not a physical "law" (itself a questionable turn of phrase). It is sloppy to say "Moore's Law".

We could call it a "rule of thumb" or a "good guess", but those are inadequate terms for an observation that has held true for 30 years and 6 orders of magnitude.

Moore's Insight? Moore's Prophecy? Moore's Unexpected But Consistent Regularity In Industrial-Economic Behavior?

Re:Punctuated Equilibrium - Phase Transition in Mo by paraax · 2006-02-20 05:08 · Score: 1

>> however there are diminishing returns past 48 processors due to communications latency

For any given problem there will be diminishing returns if you limit the size of your data set while increasing the number of processors. If you want to make use of a larger number of processors and maintain efficiency (useful computation vs communication overhead) then you need larger data sets. This won't of course help you if you need to get something done faster, it will usually help you get a more detailed answer.

Re:What's the minimum then? by DancesWithBlowTorch · 2006-02-20 05:13 · Score: 3, Interesting

In fact you can go something like 26 more levels of magnitude smaller before you start reaching the feasable limit of measurable existance. And yes, subatomic particles could theoretically be used in processors.

IANAProcessor Designer, but from what I've learned in undergraduate quantum mechanics, the problem is not the "limit of measurable existance" (I assume you are referring to the Planck Length here) but Heisenberg's uncertainty principle:

The Electrons in your transistors are "blurry". When the walls of their potential wells (i.e. the width of the wires) get to low, they will start to tunnel between them in a number that is inacceptable for the operation of a logical circuit. Note that tunneling probability is proportional to something like e to minus the potential well height, so there is no critical limit, rather a smooth transition from "no problem" to "show-stopper".

So the real question here, which is left to the audience, is at what width do we get a real problem with tunneling currents. (I assume that on contemporary CPUs, the effect is already measurable, yet correctable).

Moore's "law" by Changa_MC · 2006-02-20 05:16 · Score: 1

It's not a law, and moore didn't postulate it. It's correctly referred to as "intel's business model," since 1.5 years is a profutable schedule for new releases. When AMD first started competing seriously, Intel managed to outdo moore's law for several years.

--
Changa hates change.

Re:Moore's "law" by GuyWithLag · 2006-02-20 08:07 · Score: 1

Yes, and that's why processor prices haven't dropped in 2 years, and we're at 3Ghz-equivalent for so long....
Re:Moore's "law" by Changa_MC · 2006-02-20 13:23 · Score: 1

I'm not sure if you're agreeing with me, or sarcastically pointing out problems with what I said... Either way, your statements demonstrate that Moore's law is meaningless, which is my point. I didn't mean to imply that it was a conspiracy on the part of Intel to hamper technology, just that Moore's law gave them a good pacing mark for many years.

Moore's law held true because people believed in it, just like the 4-minute mile in running seemed unbreakable until someone did it. Now people do it all the time.

There are real limits on computer development, but they have nothing to do with Moore.

--
Changa hates change.
Re:Moore's "law" by GuyWithLag · 2006-02-20 20:28 · Score: 1

Oh, I'm actually agreeing with you - Moore's Law is nothing more than a socio-economic phenomenon that drives the technological-engineering progess in the field (and not the reverse).

On the other hand, one could argue that Intel pushed too fast ahead of the curve, thereby hitting the brick wall of current technologies before they could be replaced...

On the gripping hand, I'm just annoyed that Uploading has been postponed to after 2060, where I'll probably not make it ...

tell me about it... by Anonymous Coward · 2006-02-20 05:36 · Score: 0

my new machine isn't DRM licensed to do the "pop" assembly instruction; all I could afford was a site license of "push". Now I'm running out of stack space all the time :(

Re:tell me about it... by Captain+Zep · 2006-02-20 05:45 · Score: 1

Careful, I'm sure that running out of stack space is a patented process...
Z.

Re:What's the minimum then? by Anonymous Coward · 2006-02-20 05:40 · Score: 0

Err yeah, of course he would deny saying something stupid like that.

They better... by harris+s+newman · 2006-02-20 05:49 · Score: 0

I hear that Moore is a bad-ass. Whoever doesn't follow his law gets birdshot in the face...

Mod Parent -1, Dumb. by Anonymous Coward · 2006-02-20 06:25 · Score: 0

due to electron migration

No, they haven't. Just because a twelve year old wanna be calls it "electron migration" does not make it that.

Re:What's the minimum then? by ZonkerWilliam · 2006-02-20 06:28 · Score: 1

I am familiar with semiconductors. You mention molecular layers thick, the author references atom's thick. It's just I have heard to many people throw around the "atom's thick" with out proper reference. It's not a valid measurement. A hydrogen atom is relatively smaller than a silicon atom, at approximately ~0.12nm, while a Silicon atom is ~0.20nm,(sorry can't find the angstrom size) doesn't sound like much but at the level we are discussing,it is a large difference.

Re:Overclocking? Really? by phsdv · 2006-02-20 06:30 · Score: 1

I do not think electro-migration is caused by overclocking. Running at higher frequencies do not cause electro-migration. Probably the electro-migration was caused due to inadequate cooling...

Re:Moore's Law is so 1998 by Anonymous Coward · 2006-02-20 06:44 · Score: 0

Moore's Law is about Marketing, Getting people to buy processor after processor. Like the CD-ROM, like the .

Fucking asshole! If he hadn't said it, it wouldn't be true.

Re:What's the minimum then? by DilbertLand · 2006-02-20 06:52 · Score: 1

1 angstrom = 0.1 nanometer

Good news for silicon by Foerstner · 2006-02-20 07:24 · Score: 2, Insightful

IBM didn't invent anything new here. Rather, they proved that photolithography--the same technology used to build chips for decades--will continue to yield faster chips for the foreseeable future. In other words, silicon hasn't "hit the wall" just yet.

IBM Microelectronics doesn't have a monopoly over photolithography. They couldn't get a patent if they tried--there's prior art going back about half a century. In other words, it's good news for IBM, Intel, AMD, Texas Instruments, Micron, Freescale, Agere, Samsung, Fujitsu, and anyone else building chips.

But feel free to wave the POWER flag if you like. It's a nice architecture.

--
The US free market: two halves of a government-granted duopoly are free to set the market price.

Re:Punctuated Equilibrium - Phase Transition in Mo by Myria · 2006-02-20 07:26 · Score: 2, Insightful

Explain the human mind, then.

Simple. The amazing things that the human brain is capable of doing are parallelizable. Things like recognizing the shape of letters or phonemes in speech are definitely parallelizable tasks.

Try doing something that isn't parallelizable, like modular exponentiation of a 2048-bit number, in the human brain. It goes very slowly.

Melissa

--
"Screw Sun, cross-platform will never work. Let's move on and steal the Java language." - Visual J++ Product Manager

Re:Punctuated Equilibrium - Phase Transition in Mo by HiThere · 2006-02-20 07:33 · Score: 1

In case you hadn't noticed, the human mind is quite clumsy at numerous kinds of things. Just compare your arithmetic with that of a computer. You probably have more processor cycles/second than any computer yet available, but because you, essentially, only can use them in parallel there is a large range of problems at which computers are faster...but not infinitely faster. And we're looking ahead (one of our parallel skills) and seeing a wall.

The human mind evolved (largely) to recognize patterns in 2.5 dimensions (don't take that dimensionality too seriously, but it's larger than two and [probably] smaller than three). This is how we see things. This is how we recognize music. Etc. And it's largely how we think. Thus when we need to temporalize a pattern, we tend to compress it to the flow of a static two dimensional image, with change markers. Think musical notation, with dynamic markers. (N.B. a curved two dimensional surface remains two dimensional, but representing the curve takes a fraction of a dimension...how much depends on how complex the curve, and I can't give precise numbers even for a simple shape like a pyramid or a sphere.)

I'm speaking out of my specialty (programming), but if you doubt this, check into the literature of brain-scans of people doing mental rotations of shapes. (There was a Scientific American article on this around a decade ago, but I'm sure there is something more recent.)

--

I think we've pushed this "anyone can grow up to be president" thing too far.

used 193nm to make 30nm: factor 6.4! by phsdv · 2006-02-20 07:50 · Score: 1

I just read in http://www.eetimes.com/news/semi/showArticle.jhtml ?articleID=180204799 that IBM actually used 193nm to make the 30nm lines. An amzing 6.4 times smaller than the wave lenght used! The used a ASML machine for this. Every one (when you have enough money) could by such a machine. So no special advantage for IBM. With a little work Intel, TI, AMD, Freescale, ST or Philips could do this too.

Re:What's the minimum then? by ZonkerWilliam · 2006-02-20 07:55 · Score: 1

Okay, what I was trying for was a more accurate angstrom measurement than nanometer measurement.

Re:What's the minimum then? by krnpimpsta · 2006-02-20 08:26 · Score: 1

While the smallest chunk of silicon we could lay down would be one atom of it, there are things far smaller.

I'm confident we'll go down to the atomic level, but Silicon may have nothing to do with it. When we go down to the level of individual atoms, everything we know about Silicon will become irrelevant. In fact, even a 10nm x 10nm x 10nm block of Silicon would have a hard time performing the way we want it (yet we can already create 10nm features using STM lithography.. albiet not yet mass-produceably).

Right now, a particular patch of transistor is exists as "p type" or "n type" because it is doped with a very small amount of impurities (from 1 impurity per 1000 Si atoms (very strongly doped, ~10^19/cm^3) to 1 impurity per 1000 000 000 atoms (very lightly doped, ~10^13/cm^3)). Now, imagine what would happen if you had a feature that only contained 10 000 atoms (10nm x 10nm x 10nm). To be "very strongly doped," this patch of Silicon would have to have EXACTLY ONE impurity ion in it to behave as "strongly doped." One more or one less impurity ion would drastically change its properties. In fact, with a 1000 atom feature, you CANNOT "lightly dope it." To dope it with one atom is to already dope strongly (~10^18/cm^3). This means you would have an EXTREMELY hard time creating a traditional transistor (with 3 differently doped zones with 10nm features.

To put things in perspective, typical numbers for impurity concentrations range from 10^15/cm^3 to 10^18/cm^3. Also, I am assuming the density of Si is ~10^22/cm^-3 off the top of my head. Maybe it's ~10^23, correct me if I'm wrong.

--

New webcomic updated on Sundays: HERE

Moore's "Law" - not understood. by M0b1u5 · 2006-02-20 08:52 · Score: 1

It seems Moore's Law is just a convenient way to drive a self-fulfilling manufacturing prophecy, and its meaning is changed to suit the situation. Ikka Tuomi, quoted:

In 1965, Gordon Moore, Director of Fairchild Semiconductor's Research and Development Laboratories, wrote an article on the future development of semiconductor industry for the 35th anniversary issue of Electronics magazine. In the article, Moore noted that the complexity of minimum cost semiconductor components had doubled per year since the first prototype microchip was produced in 1959. This exponential increase in the number of components on a chip became later known as Moore's Law. In the 1980s, Moore's Law started to be described as the doubling of number of transistors on a chip every 18 months. At the beginning of the 1990s, Moore's Law became commonly interpreted as the doubling of microprocessor power every 18 months. In the 1990s, Moore's Law became widely associated with the claim that computing power at fixed cost is doubling every 18 months.

Ikka's an idiot who doesn't believe in Moore's Law by the way, but apart from that, he's a smart cookie.

--
How many escape pods are there? "NONE,SIR!" You counted them? "TWICE, SIR!"

Re:Yardsticks? I got yer yardstick. porn stars??? by seven+of+five · 2006-02-20 09:07 · Score: 1

if you're saying that the personalities of porn stars have far more data content than porno movies, then obviously you've never seen any...

Please stop mentioning Moore's Law by tod_miller · 2006-02-20 10:13 · Score: 0, Flamebait

FFS, can you drop it, it is like your dad at a disco trying to fit in - people say it like they are part of an in-crowd.

People have taken Moore's law out of context and now use it as an expression for anything that improves.

It has been changed 3 times, and get this:

It is a a nonspeculative statement about past performance, the speculative aspect was initially for 10 years, and didn't even hold until then.

can you please fuck off and never mention it again, look at any twat hack asshole writer (like those on zdnet) and all they can do is fapp themselves silly talking about how moores law might actually be the thing that is causing the speed of light to slow -

yes because they are that stupid

I hope people realise that by saying it they sound like a bearded 50 year old it geek.

*looks around* oh ffs.

--
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com

Cool - but my broadband speed still suck by Anonymous Coward · 2006-02-20 10:45 · Score: 0

nanometer this nanometer that, cpu keeps going faster and faster,
but my broadband speed is still damn slow, and getting more expensive
every year.

What's the point if we can't access the net faster?
OOOOOOOOOO dude, my CPU can crunch a gazillion flops while waiting
for yahoo or slashdot to load!!!

Awww... by StikyPad · 2006-02-20 10:50 · Score: 1

Last para. of TFA: The I.B.M. researchers performed their research on a custom piece of equipment they call Nemo, referring to the character in Jules Verne's novel "20,000 Leagues Under the Sea."

Ha! We all know what character they were really referring to.

Nemo? That's a nice name...

--
https://www.eff.org/https-everywhere

Moore's law is more vague than people give credit. by Kaldaien · 2006-02-20 11:52 · Score: 2, Insightful

Moore's law does not specify the density or even number of transistors on an integrated circuit, as many mistakenly assume; it merely states that integrated circuits double in complexity vs. cost to manufacture every 18 months. In fact, new manufacturing techniques alone, which lower the cost to manufacture can satisfy the law.

Moore's law will probably continue after quantum well transistors are implemented and minituarized. The Cell architecture and push for multi-core processors lend themselves well to Moore's law as well. I would wager designing 4-8 core CPUs, multi-core CPUs with shared caches and the new AMD chips that integrate the memory controller rather than using a Northbridge easily satisfy Moore's law.

Re:Punctuated Equilibrium - Phase Transition in Mo by mlyle · 2006-02-20 14:40 · Score: 1

Try doing something that isn't parallelizable, like modular exponentiation of a 2048-bit number, in the human brain. It goes very slowly.

Bad analogy.

You might want to think of something better than modular exponentiation to use as your example (e.g. there's several well-known parallel algorithms for doing just that).

Also, the human mind is bad at doing these things because nearly everyone operates on numbers symbolically-- the raw computing hardware underneath is not used to do the math, but it's rather like the math is done in a very clunky interpreter with only a fraction of system resources.

RIT's Lithography leads the way by banana6986 · 2006-02-20 16:44 · Score: 1

http://www.microe.rit.edu/research/lithography/res earch/immersion.htm and no not every student at RIT hates it (comment posted earlier)....actually not that many do...just the antisocial people that sit in their room all day. I pay for the education and reputation....as you can see we do have a very good microe program which im proudly part of. Enjoy the link...

Cold start time by tepples · 2006-02-20 17:39 · Score: 1

Besides, the only people who still think yesterday's systems were more responsive or efficient are the ones who haven't revisited them.

Why does a PlayStation 2 game boot in 30 seconds while a Super NES game boots in 3?

Re:Cold start time by Just+Some+Guy · 2006-02-21 03:15 · Score: 1
Why does a PlayStation 2 game boot in 30 seconds while a Super NES game boots in 3?
Because the SNES booted like this:
1. Clear the screen
2. Start loading the game
while the PS2 boots like this:
1. Run hardware diagnostics
2. Initialize all components of a complex computer system
3. Start loading the game
In other news, solid-state cartridges have less latency than optical drives.
--
Dewey, what part of this looks like authorities should be involved?

nothing fishy by tepples · 2006-02-20 17:43 · Score: 1

Ha! We all know what character they were really referring to.

Something to do with slumberland?

bits are not bytes by gracefool · 2006-02-21 12:11 · Score: 1

If only journos would learn that bits are not the same as bytes. From the article:

Currently, the densest computer memory chips store 4 billion bits of information

Ah, that's 476Mb. I think they mean bytes. To be precise, not 4 billion bytes but 4 gibibytes.

Re:bits are not bytes by Anonymous Coward · 2006-02-21 13:46 · Score: 0

lol, you mean 476MB

Slashdot Mirror

Moore's Law Staying Strong Through 30nm

199 comments